cover_letter / art_GY_Sfm1sDYs

Dear Robinhood Hiring Team, Robinhood's mission to democratize finance sits at the center of one of the most consequential economic shifts in modern history — the $124 trillion intergenerational wealth transfer that will reshape who participates in financial markets. That mission resonates with me directly: I built Fintellect AI precisely to close the financial literacy gap that keeps younger, first-time investors on the sidelines. Moving from founder to Staff PM at Robinhood means working at the infrastructure layer that makes that democratization safe and durable — and that is where I want to apply the next chapter of my career. **Technical and AI/ML Foundation** My technical credibility spans the full stack from ML research to production platform engineering. In 2004 I hand-coded backpropagation-through-time in C++ for a protein structure prediction system that eventually became a NeurIPS-published paper; in 2026 I rewrote that system in PyTorch across five architectures (feedforward through Transformer and ESM-2), scaling from 413 to 8 billion parameters. That arc — from first principles to modern deep learning — informs how I evaluate AI automation claims and make build/buy/integrate decisions. More directly relevant to this role: I built aeval, a local-first model evaluation platform with five core eval types (factuality, reasoning, instruction-following, safety, code generation), adversarial safety testing with refusal detection, bootstrap confidence intervals, Welch's t-test, and Cohen's d effect sizing — the statistical rigor needed to trust an ML model before it touches a production decision. I also built an RL post-training workbench that benchmarks 12 algorithms (PPO, GRPO, DPO, DAPO, REINFORCE, RLOO, SimPO, IPO, KTO, ORPO, SPPO, REINFORCE++) across TRL, VeRL, OpenRLHF, and NeMo RL with live SSE metric streaming. These projects were not academic exercises — they were built to answer the same question Robinhood's Operations Platform team faces: *can I trust this model's output before it routes a real case?* On the platform infrastructure side, at Intuit I owned the ICE developer platform that scaled to 675M+ engagements in FY23 across QuickBooks, TurboTax, Mint, Mailchimp, and Credit Karma, growing 275% YoY. I drove a throughput migration from 6K to 50K TPS via rSocket, supporting approximately 1.5M concurrent connections at sub-25ms TP99. I also delivered the ICE Self-Service platform — DevPortal, GitOps config, ICE Playground — reducing developer onboarding from two to three weeks down to minutes in pre-production and under 24 hours for production, while mitigating over $1M in projected opex growth. Scaling foundational platform services under strict reliability and latency constraints is work I have done before. **Why This Role** The Operations Platform role at Robinhood sits exactly at the intersection where I have spent my career: platform infrastructure that enables human operators, ML models that automate high-volume decisioning, and customer-facing flows that reduce friction without compromising safety. Owning the roadmap across case management, ML alert resolution, new-product readiness, and the customer fraud-control experience is the kind of multi-dimensional, high-stakes product problem I am wired to solve. What specifically excites me is the AI-based automation of fraud alert resolution. This is not a generic ML integration — it requires the same evaluation discipline I built into aeval (safety gates, regression detection, confidence thresholds) applied to decisions with real financial consequences for real customers. Getting that wrong in either direction — false positives that freeze legitimate accounts, false negatives that let fraud through — has direct, measurable impact on the trust metrics this team owns. I have the technical depth to partner credibly with data science on model design and the PM discipline to define the acceptance criteria, monitoring strategy, and rollback conditions that make deployment responsible. **Selected Prior Experience** - **Intuit — ICE Self-Service Platform:** Delivered DevPortal, GitOps config, and ICE Playground, reducing developer onboarding from 2–3 weeks to minutes in pre-prod and under 24 hours for production; mitigated $1M+ in projected opex growth — directly analogous to the Opex reduction mandate in this role. - **Intuit — Platform Scale:** Achieved 275% YoY growth in ICE engagements, scaling to 675M+ in FY23; drove rSocket migration from 6K to 50K TPS supporting ~1.5M concurrent connections at sub-25ms TP99. - **Intuit — Drift Detection Program:** Initiated MSaaS Drift Detection and Resolution — wrote a Java JAR library to scan Git repos for configuration drift, partnered with Design on DevPortal UI, and built a remediation roadmap using OpenRewrite. Systematic detection of configuration risk at scale is directly transferable to fraud-control system integrity. - **aeval — AI Model Evaluation Platform:** Built adversarial safety testing, refusal detection, bootstrap confidence intervals, and CI/CD regression gates — the evaluation infrastructure needed before deploying ML models to fraud alert resolution. - **RL Workbench:** Benchmarked 12 RL algorithms across TRL, VeRL, OpenRLHF, and NeMo RL with standardized throughput, memory, and convergence metrics — establishes the technical credibility to partner with data science on model selection and post-training strategy for alert automation. - **Splunk — Search Orchestration:** Owned Search Service (Go microservices), Search Catalog (PostgreSQL metadata), and SPL/SPL2; delivered Scheduler Service end-to-end in four months; led query performance optimization achieving up to 10x improvements for a Fortune 500 beta customer. Case management and investigation tooling at scale is a direct analog. - **Kaiser Permanente — Logging-as-a-Service:** Led enterprise rollout of Splunk Logging-as-a-Service at 1.7 TB daily volume across 200+ internal customers; built Redis-based caching for scalability and fault tolerance. Operating high-volume, high-reliability internal platforms for enterprise-scale operations teams is the foundation this role requires. **Closing** Robinhood's mission only holds if the platform underneath it is trustworthy — if fraud controls protect customers without becoming the friction that drives them away, and if the ML systems making those decisions are evaluated rigorously enough to deploy with confidence. That is the problem this role exists to solve, and it is the problem I have been building toward across platform infrastructure, AI tooling, and developer-facing product work for over a decade. I would welcome the opportunity to discuss how my background maps to the specific challenges your Operations Platform team is navigating. --- O. Felix Amoruwa famoruwa@berkeley.edu · 909-731-9011 · felixamoruwa.info